OpenAI is developing a long-term system to predict if a user is over or under 18. Anyone identified as under 18 will be directed to an age-appropriate ChatGPT experience which will include blocking graphic sexual content and, in rare cases of acute distress, potential law enforcement involvement. The company said that when its principles conflict, it prioritizes teen safety over privacy and freedom.
The AI firm will also release new parental controls by the end of the month. Parents will be able to link their account with their teen’s (ages 13+) account via an email invitation. These controls will allow parents to manage how ChatGPT responds to their teen. Parents will also be able to disable features like memory and chat history. Parents that enable it will get notifications if their teen is in distress, and there’s a control to set blackout hours to restrict usage.
OpenAI has acknowledged that getting age prediction right is difficult. The company’s solution involves defaulting to the safer, under-18 experience when unsure of a user’s age. Adults will have a way to prove their age, but the firm didn’t specify what that process entails. The company says the approach to this matter has been guided by conversations with experts, advocacy groups, and policy makers.
The announcement is part of a broader effort to improve the safety of all users on ChatGPT. The company said it’s still continuing to learn and improve its systems based on feedback it gets from partners and experts. Ensuring safe access to ChatGPT is probably a wise move on the part of OpenAI, as it’s quite a new technology, and we don’t yet know the long-term effects it could have on children. Also, against the backdrop of governments requiring age verification on adult websites, it makes sense for OpenAI to be proactive in this regard.
Source: OpenAI